Street Trees and Summer Cooling: What the Data Can and Cannot Support
A council planted 4,200 street trees across three districts between 2015 and 2019. A 2023 survey measured summer afternoon surface temperatures: tree-lined streets averaged 3.1°C cooler than matched streets with no trees. Cooling was greatest in Northfield, which has the densest canopy (4.4°C), and smallest in Eastgate, where trees were planted in narrow pits with little soil (1.6°C). Residents of tree-lined streets reported 22% more walking. The same streets carried 12% more shade-tolerant weeds. Districts were not randomly assigned: Northfield is also the wealthiest district and received traffic-calming measures in 2016. The report notes that 9% of planted trees had died by 2023, mostly in Eastgate.
1. Which one of the following is best supported by the passage? A. Street trees cause residents to walk more. B. The size of the cooling effect depends on how the trees were planted. C. Replanting Eastgate's trees would produce 4.4°C of cooling. D. Shade-tolerant weeds are caused by street trees.
Read the worked answer
Answer: B.
- B is a modest, comparative claim: the effect ranged from 1.6°C to 4.4°C across districts, and the passage explicitly links the weakest effect to narrow pits with little soil. That is supportable as an association.
- A and D are causal claims about walking and weeds. The passage gives only co-occurrence (22% more walking, 12% more weeds); no mechanism or controlled comparison is offered.
- C transfers Northfield's figure to Eastgate and assumes the densest canopy could be recreated - the passage says Eastgate's trees had the highest mortality and the poorest planting conditions.
Note the trap in B: it says 'depends on', not 'is caused only by'. The passage supports a dependency even though confounding exists.
2. The council leader says: 'This proves that planting trees will cool every street by 3.1°C.' Identify two features of the report that make this an overstatement, and explain each.
Read the worked answer
1. The effect is heterogeneous, not uniform. 3.1°C is a district-wide average spanning 1.6°C to 4.4°C. Applying the mean to 'every street' ignores the eight-fold-ish spread and the fact that Eastgate - the worst-planted district - got less than half the average.
2. Allocation was not random, so causation is unproven. Northfield is both the wealthiest district and the recipient of 2016 traffic-calming measures. Either could contribute to cooler streets. A correlation across self-selected districts cannot 'prove' that trees did the work.
A third acceptable point: 9% tree mortality by 2023 means the canopy measured is not the canopy that will exist, so the figure is not a stable prediction.
3. What single piece of additional evidence would most strengthen the claim that the trees caused the cooling? Justify your choice against at least one alternative.
Read the worked answer
Best: before-and-after temperature data for the same streets, compared with matched control streets over the same period.
Why this works: showing that tree-lined streets cooled relative to their own baseline while control streets did not removes stable between-street differences (aspect, building density, wealth) that could explain the 2023 gap. It converts a cross-sectional comparison into a difference-in-differences comparison.
Why not simply 'more districts'? Adding districts increases sample size but repeats the same confounding structure - the districts would still be non-randomly assigned. Why not 'more accurate thermometers'? Measurement error is not the weak link here; study design is. Why not 'survey residents'? Self-report about walking says nothing about surface temperature.